banks™use of credit derivatives and loan pricing: what is ... · is associated with...
TRANSCRIPT
Banks�Use of Credit Derivatives and Loan Pricing:
What Is the Channel and Does It Persist Under
Adverse Economic Conditions?�
Lars Nordeny, Consuelo Silva Bustonzand Wolf Wagnerx
DRAFT. DO NOT CITE WITHOUT PERMISSION OF AUTHORS.
Abstract
This paper studies whether the use of credit derivatives at banks has an impact
on the spreads banks charge to their corporate borrowers, and if so, through which
channel(s) this occurs. We �nd that a bank�s gross position in credit derivatives
is associated with signi�cantly lower loan spreads in syndicated lending markets,
while the bank�s net position is not related to loan spreads. We argue that this is
consistent with banks passing on risk management bene�ts to their corporate bor-
rowers, but not with other channels through which credit derivatives may a¤ect
loan pricing. We also �nd that the bene�ts increase for borrowers that are more
likely to be actively traded in the credit derivatives market. The evidence further
indicates that risk management using credit derivatives continues throughout the
crisis of 2007-2009 since i) the bene�t of borrowing at lower spreads from credit-
derivative active banks does not fall throughout the crisis, ii) active banks have
persistently lower loan charge-o¤s than other banks, iii) active banks cut lending
less than other banks during the crisis. Taken together the evidence highlights
important risk management bene�ts from �nancial innovations that persist under
adverse conditions �that is, when they matter most.
Key words: Financial innovations, credit derivatives, syndicated loans, loan
pricing, �nancial crisis
�We thank seminar participants at Tilburg University for comments.yRotterdam School of Management, Erasmus University and ERIM. E-mail: [email protected], European Banking Center, and Department of Economics, Tilburg University. E-mail:
[email protected], European Banking Center, TILEC, and Department of Economics, Tilburg University.
E-mail: [email protected].
�Credit derivatives [have] contributed to the stability of the banking system by allowing
banks . . . to measure and manage their credit risks more e¤ectively. . . .�
Alan Greenspan, 2005
"The boom in subprime mortgage lending was only a part of a much broader credit
boom characterized by ... the creation of complex and opaque �nancial instruments
that proved fragile under stress."
Ben Bernanke, 2008
1 Introduction
Financial innovations are at the centre of an intense debate on how to shape the future
global �nancial system. The dominant view prior to the crisis of 2007-2009 was that
�nancial innovations are bene�cial for the �nancial system. The experience of the crisis
has led to an �at least partial �reassessment of this view. Some policy makers have
even adopted the opinion that the use of �nancial innovations needs to be restricted
or prohibited. In general there is an emerging notion that �nancial innovations, while
bene�cial under favorable conditions, can aggravate �nancial crises. Whether the latter
is the case is likely to depend on how these innovations are used: �nancial innovations
may for example be employed by �nancial institutions to control their risks but may
also make these institutions reliant on their usage and encourage them to increase their
risk-taking. Despite the importance of this issue, there is still relatively little evidence
on the channels through which �nancial innovations a¤ect �nancial institutions and
how this channel a¤ected institutions during the crisis. This paper examines whether
and through which channel credit derivatives in�uence bank behavior in the lending
market and how this has a¤ected banks during the crisis.
Credit derivatives are probably the most signi�cant �nancial innovation of the re-
cent decade and banks are major buyers and sellers of protection. Unlike traditional
debt instruments (such as bonds and loans) credit derivatives make it easy to hedge
or source credit risk of a single borrower or a pool of debtors. The most prominent
credit derivative is the credit default swap (CDS). A CDS is a contract under which a
protection buyer makes periodic payments to a protection seller in exchange for pro-
tection against the default of a reference entity. The market for credit derivatives has
grown dramatically during the last decade. The Bank for International Settlements
(BIS) estimates that the market peaked in 2007 with an outstanding amount of $60
trillion. Notional amounts remained very high after the onset of the crisis but following
the failure of Lehman Brothers declined to $41 trillion at the end of 2008 and to $31
1
trillion in June 2010. However, in contrast to other credit markets the CDS market did
not break down during the crisis.
Studies that examined banks�use of �nancial innovations show that prior to the
crisis these instruments lead to large credit extension extensions and more favorable
lending conditions. In particular, there is evidence that loan sales (Cebenoyan and
Strahan (2004)), Collateralized Debt Obligations (Franke and Krahnen (2005)) and
Collateralized Loan Obligations (Goderis et al. (2006)) are associated with an increase
in lending at banks. Lower spreads are associated with loans intended for subsequent
sale (Guener (2006) or securitization (Nadauld and Weisbach (2010)). Hirtle (2009)
�nds that there is a positive link between a bank�s net position in credit derivatives and
loan spreads. In contrast, Ashcraft and Santos (2009) document that �rms face higher
loan spreads after they start being traded in the CDS market. They argue that this
e¤ect is driven by reduced incentives for banks to monitor �rms.
This studies considers credit derivative usage at U.S. banks and their impact on the
banks�pricing of syndicated loans and the characteristics of the commercial bank lend-
ing in general. Our main data sources are the LPC DealScan database which provides
detailed loan-level information and the Call Reports which contain bank-level informa-
tion. We consider four di¤erent channels through which credit derivative usage a¤ects
loan pricing. Brie�y, derivatives provide bene�ts that can be passed on to borrowers
if banks use them to hedge credit risk, to reduce economic or regulatory capital or to
manage credit risk. They can also cause ine¢ ciencies if the transfer of risk leads to
incentive problems that results in lower screening and monitoring of borrowers. For
each channel we derive predictions about how it can be empirically identi�ed. The key
predictions turns out to be that the risk management can be operative without banks
taken on a net position in credit derivatives: banks can reduce concentrated exposures
by buying protection but at the same time source credit risks on underrepresented (or
absent) borrowers by buying protection. All other channels operate through banks
taking a net position in credit derivatives.
Our principal result is that, after controlling for lender, loan and bank characteris-
tics, banks�gross positions in credit derivatives are negatively and signi�cantly related
to the (loan) spreads they charge for the average corporate loan. By contrast, net po-
sitions do not display any association with loan spreads. This provides support for the
risk management channel but not for the other channels through which credit deriv-
atives may a¤ect loan pricing. The e¤ect is very robust. In particular, it survives
when we take into account various simultaneity and endogeneity concerns. The e¤ect,
however, increases in magnitude if we consider �rms that are most likely to be actively
2
traded in credit derivatives markets. This is to be expected as banks should �nd it eas-
ier to manage the risks of these borrowers. The largest e¤ect obtains for �rms that are
rated investment-grade: the estimates imply that a one-standard deviation increase in
the banks�gross credit derivative position lowers their loan spread by 18% (44 bps). We
also �nd that the risk management bene�ts also extend to �rms that are unlikely to be
traded in the credit derivative market: their spread falls by 4% (10 bps). This suggests
that risk management bene�ts are passed on to the entire portfolio of borrowers and not
only the ones that can be directly traded (this may either be because of pseudo-hedging
were a bank uses a correlated traded exposure to hedge risk on an untraded exposure
or because diversi�cation bene�ts lower risk premia required for lending).
We also analyze whether our main result holds during the crisis of 2007-2009. If
banks use of credit derivative is e¤ective in managing risk, we would expect the advan-
tage relative to other banks to persist. We �nd that loan spread increased for all banks
during the crisis consistent with the fact that this crisis was driven by systemic factors
that are not easily diversi�able. However, consistent with e¤ective risk management
we �nd that banks active in credit derivatives still charge spreads that are lower than
other bank �the spread di¤erence in fact does not fall compared to pre-crisis levels.
Finally, we investigate the e¤ects of credit derivative usage on banks lending and
bank performance during the crisis. Risk management suggests that active banks should
lower the variability of their cash-�ows and are hence less likely be end up constrained
(Froot, Scharfstein and Stein (1993)). Consistent with this we �nd that active banks
cut lending back by signi�cantly less than other banks. These banks also do not seem
be generally more aggressive as their pre-crisis lending levels is comparable to other
banks. Furthermore, we expect banks that actively manage their credit risks to have
lower losses and not to su¤er di¤erently from the �nancial crisis than the other banks.
Accordingly, we �nd that banks with a larger gross CDS position have lower charge-o¤s
than other banks and this advantage was not eroded (even partially) in the crisis. This
parallels our results from the loan pricing regressions.1
Taken together the analysis provides consistent and strong evidence that bank use
credit derivatives to actively manage their risk.2 There is no evidence for other channels
through which credit derivatives may a¤ect loan spreads. Corporate lenders bene�t
1We also �nd that during the entire sample period loan spreads of banks active in credit derivative
markets are much lower than at peer banks. This further speaks to risk management resulting in more
stable lending.2Our results are complementary to the recent evidence on the link between risk management, control
and performance of US bank holding companies (Ellul and Yerramilli, 2010).
3
from risk management through lower spreads and these bene�ts do not seem to be
limited to the borrowers to which the �nancial innovation directly applies. Our results
also show that in the case of credit derivatives the bene�ts also extend to crisis period.
These bene�ts do not only arise through more favorable lending conditions but also
through a more stable supply of credit. Hence, credit derivatives still extend bene�ts
when they are most needed. All in all our results provide a positive message for the
bene�ts of �nancial innovation � even in circumstances in which the markets where
these innovations are traded are under great stress.
The remainder of the paper is organized as follows. Section 2 develops a set of
hypotheses in order to identify the channel through which credit derivatives might
a¤ect corporate loan spreads. Section 3 describes the data. Section 4 explains the
empirical strategy and presents the results. Section 5 concludes.
2 Hypotheses
Academics and practitioners have suggested di¤erent channels through which credit
derivatives (and risk transfer activities in general) can a¤ect bank lending. In this
section we brie�y summarize key channels. We also explain our approach for how to
identify these channels empirically.
Credit derivatives allow banks to transfer the risk of their portfolios to other entities,
either by selling loans or by hedging the exposure via purchasing protection. This may
reduce their incentives to screen and monitor borrowers (e.g., Morrison (2005)). We
refer to this as the Incentives Channel. Ashcraft and Santos (2009) provide evidence
for this channel. Ashcraft and Santos investigate the e¤ect of the �rm being traded in
the CDS market on the (loan) spreads they have to pay for their corporate loans. They
argue that once a �rm is traded in the CDS, banks can hedge its exposure to this �rm.
This may lower banks�incentives to monitor. The �rm�s borrowing cost should then
increase �as it becomes riskier. Consistent with this, Ashcraft and Santos �nd that
informationally opaque �rms, who bene�t the most from bank monitoring, face higher
spreads after the onset of trading in the CDS market.3
Credit derivatives may also a¤ect bank lending through the Risk Management Chan-
nel. According to this channel credit derivatives allow banks to better manage the credit
risk of their portfolios. They can buy protection on overrepresented exposures and sell
protection on credits they have little exposure to. Banks can also use credit derivatives
3Marsh (2000) �nds that the announcement e¤ect of a new bank loan is weakened when a bank
actively uses securitization techniques to transfer of risk �consistent with weakened bank incentives.
4
to keep the overall risk of their portfolio close to the target level. Among others, this
provides bene�ts as it reduces the likelihood of �nancing constraints becoming binding
(Froot, Scharfstein and Stein (1993)). Risk management bene�ts may also obtain indi-
rectly: the use of credit derivatives may induce banks to more rigorously measure their
credit risks. An increased awareness of risks may make banks more e¢ cient in their
lending behavior. Empirical research provides evidence that risk management bene�ts
enables banks to extend larger loan volumes (Goderis et al., (2007), Franke and Krah-
nen (2005)) or to pass on the bene�ts to their borrowers through lower spreads (see
Cebenoyan and Strahan (2004) for loan sales). If this channel is operative, we would
expect banks that are actively trading credit derivatives to reduce the interest rate
charged to borrowers. Hirtle (2009) examines this hypothesis. Controlling for bank
and loan characteristics, she �nds that for large borrowers the net position of credit
derivatives held by banks has a negative e¤ect on loan spreads, and argues that this
�nding is consistent with banks managing credit risk.
There are two additional channels through which credit derivatives may in�uence
loan pricing. Both channels suggest a negative e¤ect on loan spreads. According to the
Hedging Channel banks hedge their exposures by purchasing protection in derivatives
markets, which enables them to take on new loans. Nadauld and Weisbach (2011) study
whether this channel has an impact on loan pricing. Nadauld and Weisbach examine
the spreads of loans that are subsequently securitized. They �nd that loans that were
later included in a CLO exhibit lower spreads when they are issued. Another channel,
closely related to the hedging channel, is the Capital Relief Channel. This channel is
based on the idea that bank lending is constrained due to insu¢ cient regulatory capital.
Credit derivatives can be used to alleviate this constraint by buying protection from
third parties, thus releasing equity for new lending. This allows banks to grant new
loans and price loans more aggressively. Broadly consistent with this channel, Loutskina
and Strahan (2006) show that securitization diminishes the impact of bank �nancial
condition on loan supply.
While most of the studies have focused on one channel, our paper considers these
channels jointly and aims to identify the key channel(s) through which credit deriva-
tives in�uence corporate loan spreads. It should be noted that the channels vary with
their prediction regarding the impact on loan spreads (a spread reduction is suggested
by the risk management, hedging and capital relief channel, while a spread increase is
consistent with the incentive channel). However, the key innovation in our paper that
ultimately allows us to identify the dominant channel is that we separately consider the
e¤ect of the gross and the net credit derivative position on spreads. We argue that all
5
Table 1: Predictions of the e¤ect of di¤erent CRT channels on Spreads
Hypothesis Net CD Position Gross CD Position
Incentives Channel (+) No e¤ect
Hedging Channel (-) No e¤ect
Capital Relief Channel (-) No e¤ect
Risk Management Channel No e¤ect (-)
channels except the risk management channel require the bank to take a net-position
in the credit derivative market (that is, to sell protection). Under the hedging channel
risk is only reduced if the bank sheds risk net, that is, buys more protection than it
sell. Similarly, under a capital relief is only provided if the bank overall reduces its
risk, again requiring the bank to take a net-position. Finally, the incentive channel also
requires banks to sell risk �but not to buy. The only channel that can become operative
without a net position is the risk management channel. For example, diversifying the
portfolio by shedding risk on overrepresented borrowers and assuming risk on under-
represented exposures can be achieved without taking a net position. Improvement of
the measurement of risks requires regular use of credit derivatives but not to take a net
position. We thus argue that �nding an association between gross positions and loan
spreads supports the risk management channel. In addition, absence of a relationship
between the net-position and the spread is evidence against the presence of each of the
three other channels.
Table 1 summarizes the predictions of the various channels for loan spreads, and
whether the relationship comes through the net or the gross position. Note that we
cannot distinguish between the hedging and the capital relief channel �but between all
other channels.
3 The data
Our analysis is based on individual loan transaction data from the LPC DealScan
database and bank level data from the US Call Reports. From the �rst database we
obtain information on loan characteristics of syndicated loans, such as loan spread over
LIBOR, loan maturity, loan amount, currency, loan purpose, loan type. We also obtain
borrower characteristics such as industry, sales, rating, stock market listing. We only
consider completed term loan transactions. The database also provides information
about the lead arrangers that are involved in the syndicate. We restrict ourselves to
6
loans with a single lead arranger as in the case of multiple lead arrangers it is di¢ cult
to attribute credit derivative use of individual banks to the syndicate group. We match
the lead arranger with bank-level data from the Call Reports. From the Call Reports we
obtain quarterly bank balance sheet and income statement information. We also collect
information about banks�o¤-balance sheet activities from these reports. From this we
construct our main variables of interest: the outstanding volume of credit derivatives
purchased and sold by the bank in each quarter. The sample covers the period from
the �rst quarter of 1987 until the �rst quarter of 2010 and comprises a total of 7,959
loan observations and 316 banks.
Table 2 reports summary statistics for our sample. The average (all-in) loan spread
in our sample is 238.18 basis points and varies between 30 and 455 basis points. Our
main variables of interest are banks�gross and net credit derivative positions. The gross
position (the outstanding sum of protection bought and sold) is on average around
45% of total assets. The net position (the di¤erence of outstanding sold and bought
protection) is only 2% of assets on average (but varies widely between banks). Figure
1a and 1b depict the evolution of the quarterly averages of the gross and net credit
derivatives positions over time4. It can be seen that, starting from the �rst quarter
of 1996 (when reporting requirement for credit derivatives start), the gross position
held by banks increases over time. The net position �uctuates between -0.1% and 4%
of assets. We can see that starting from the end of 2005 banks increased their net
purchase of protection, presumably in anticipation of a higher share of problem loans.
Figure 2 compares the loan spreads charged by banks that are active in credit
derivatives markets with the ones of banks that are not. For this �gure we consider a
bank being �active�from the moment on it either purchases or sells protection for the
�rst time. We can see that throughout the sample period active banks tend to charge
lower spreads than passive banks.5 The mean di¤erence in the spread of active and
passive banks is 46 bps and this di¤erence is very signi�cant (t-statistic of 10.1). We
also note that during the sample period there does not seem to be any trend in the spread
di¤erences among the group of banks. This is �rst evidence for credit derivatives use
being associated with a persistently lower loan spread. In addition, the spreads of the
active banks seem to be more stable over time compared to their passive counterparts,
consistent with risk management e¤ects.
4These �gures exclude the Bank of America, which bought very large amounts of protection in 2005
and 2007.5In the �gure for some quarters averages for passive banks are missing since there were no loans
originated by these banks.
7
4 Empirical method and results
4.1 The empirical strategy
We estimate a loan-spread model that controls for loan, borrower and bank character-
istics. We proxy banks�credit derivative use with the gross and net positions of credit
derivatives scaled by assets. A signi�cant negative relationship between the gross posi-
tion and the loan spread supports the risk management channel. A negative signi�cant
coe¢ cient on the net position would provide evidence for the hedging or capital re-
lief channel, while a positive relationship would be consistent with credit derivatives
leading to incentive problems. The various channels also lead us to expect that the
impact of credit derivative use may depend on the borrower type and whether banks
operate under adverse circumstances. In a second step we hence also study whether the
loan-spread impact di¤ers among borrowers and whether it changes during the crisis of
2007-2009.
In order to investigate whether credit derivative use has an e¤ect on loans spreads
we estimate the following model at the loan-level:
spreadb;f;l;t = �+ �1bankb + �2yeart + �3grossCDb;t + �4netCDb;t +KXi=1
�iFi;f;t
+KXi=1
iLi;b;f;l;t +KXi=1
�iBi;b;t + �b;f;l;t; (1)
where b denotes the bank, f the borrower (�rm), l the loan and t time. In (1) spread is
the loan spread, bank is a set of bank dummies and year is a set of time dummies. The
term grossCD denotes the sum of credit protection sold and purchased by a bank and
netCD is the di¤erence between credit protection purchased and credit protection sold.
The terms Fi are borrower characteristics. These characteristics include dummies indi-
cating the industry group of the borrower. We also include the logarithm of the sales
in US dollars. We expect �rms with more sales to have lower spreads since large �rms
are more likely to have built a reputation and are less likely to su¤er from problems
of informational asymmetries. We also include dummies indicating whether the com-
mercial paper of the borrower is rated (rating) and whether the borrower is listed on
the stock market (ticker). We expect a negative association between the dummies for
rating and stock market listing and the loan spread because rated and public �rms are
likely to face lower informational asymmetries. Further we control for a set of dummies
that indicate the S&P senior debt rating of the borrower (using BBB as the omitted
category). Within the set of ratings, we expect higher rated �rms to be charged lower
8
spreads.
The vector L refers to loan characteristics. Following Harjoto (2006), these controls
include two dummy variables that indicate whether the database denotes a loan as
secured and whether it denotes a loan as unsecured (the omitted category are loans for
which securitization information is missing). It is not clear what sign to expect for these
dummies. Safe borrowers may use collateral to signal their type to the lender (Besanko
and Thakor, 1987; Chan and Kanatas, 1985). If this is the case, secured loans should
be associated with lower spreads. However, there is evidence suggesting that lenders
require collateral for riskier borrowers, which would lead to higher spreads (Berger and
Udell, 1990; Berger, Frame and Ioannidou, 2011). We also include among controls the
logarithm of the loan amount in US dollars (log(amount)). Again, the loan amount
coe¢ cient can be positive or negative. Larger and safer �rms usually demand larger
loans, hence we should expect lower spreads for such loans. However, larger loans have
also a higher probability of default and may result in overexposures in banks�credit
portfolios, suggesting higher spreads. The next set of variables contains dummies for
the loan maturity: shortmaturity for term loans with maturity of less than two years,
intermediatematurity for term loans with maturity between two and �ve years, and
longmaturity for term loans with a maturity exceeding �ve years (we will use short
maturity as the omitted category). The expected sign on these dummies is ambiguous
as well. There is some evidence of longer maturity loans being associated with higher
spreads (Dennis, Nandy and Sharpe, 2000) but other studies show that short maturity
loans exhibit higher spreads (Strahan, 1999). We further include a set of loan purpose
dummies (corporatepurposes, acquisitions, backupline, and debtrepayment). Finally,
we consider dummies for the tranche type. TERM indicates terms loans without
a tranche structure and TERMA, TERMB; TERMC+ indicate whether a loan is
designated to tranche A, B, C or higher, respectively.
The terms Bi stand for bank characteristics. We include as a proxy for bank size the
logarithm of assets. We expect this coe¢ cient to be negative given that larger banks
are expected to have a lower cost of funds due to better access to debt markets. We
also include a measure of a bank�s liquidity equal to cash plus securities over assets,
liquid Ass:=TA. We expect this coe¢ cient to be negative as well, re�ecting that liquid
banks �nd it cheaper to fund loans. Further we include as additional controls the return
on assets (ROA), the amount of charge-o¤s over assets (chargeoff=TA), subordinated
debt over assets (subdebt=TA) and equity over assets (equity=TA).
9
4.2 Credit derivative use and loan spreads
Table 3 reports the results of regressions that relate loan spreads to bank credit deriva-
tive positions. All regressions include borrower controls, loan controls and industry, loan
purpose and year dummies. Standard errors are clustered at the bank level. Regression
1 includes next to the gross and the net positions the bank controls. The coe¢ cient
of the gross position takes a negative value (-9.36) and is signi�cant at the 1%-level.
The coe¢ cient of the net position is not signi�cant. This provides support for the risk
management channel but not for the other channels. The magnitude of the e¤ect for the
gross position indicates economic signi�cance. It implies that a one standard-deviation
increase in the ratio of the gross position over (total) assets decreases loan spreads by
about 9 basis points. Given a mean spread of 238 bps this implies spreads fall on av-
erage by 4%.# 6 This is a considerable impact �in particular since this is the impact
on the average borrower in the syndicated loan market (many of these borrowers will
not be actively traded in the credit derivatives market). It should also be noted that
our coe¢ cient likely understates the risk management e¤ect of credit derivatives. This
is because our data on credit derivatives also includes credit derivatives for trading
purposes.7 We do not have information about the share of credit derivatives used for
hedging in our sample. However �surveying a sample of banks trading credit deriv-
atives �BBA (2006) found that about 50% of the protection purchased by surveyed
banks was held for hedging. This suggests that the true impact of hedging derivatives
is more in the magnitude of 18 bps rather than 9 bps.
Among the borrower controls, we can see that larger �rms are charged lower spreads.
The same is found for rated �rms and �rms which have a stock exchange listing. Various
rating category dummies turn out to be signi�cant as well (the insigni�cance of the other
rating dummies is due to the fact that for these ratings there are only few observations).
Among the signi�cant rating categories, loan spreads are found to decline with the �rm�s
S&P rating �as expected. Turning to the loan controls, we �nd that there is a negative
and signi�cant association between loan amount and loan spreads. This may re�ect
the tendency for large loans to be given to larger, established, �rms. Secured loans
have signi�cantly higher, and unsecured loans have signi�cantly lower, spreads. This is
explained by banks more likely requiring collateral for lending to risky �rms. Among the
maturity variables, the long-term maturity dummy enters with a negative sign and is
6Given an average loan size of # U$, the implied yearly savings for borrowers are # U$ for the
average loan.7Although the Call Reports distinguish derivatives held for trading for other derivatives, they do
not do this for credit derivatives.
10
weakly signi�cant (at the 10% level). Finally, all the loan tranche indicators are positive
and signi�cant. Since the omitted category is loans without a tranche structure, this
indicates that tranched loans are more risky and consequently command higher spreads.
From the bank controls only the charge-o¤s are signi�cant. They enter with a positive
sign. This likely re�ects that banks that have many problem loans book incur higher
costs and pass these costs on to their borrowers.
Regression 2 includes bank �xed e¤ects instead of bank controls. The coe¢ cient
on the gross position increases in absolute value to -10.93. The net position remains
insigni�cant. The other coe¢ cients in the model are mostly unchanged. We take this
model to be our baseline model. There is the concern that the insigni�cance of the
net position is driven by a potential multicollinearity between net and gross positions.
However, the correlation among these variables is not very high (0.22). To be sure,
regression 3 reports results where the gross position is excluded. The net position re-
mains insigni�cant. Some of the previous results suggest that loan characteristics might
be jointly determined with the loan spreads. In regression 4 we follow the literature by
estimating a model that excludes the loan controls. The coe¢ cient of the gross position
now increases in absolute value to -14.58. This surely re�ects that some of the loan
controls are correlated with credit derivative use at banks. However, the coe¢ cient on
the gross position remains signi�cant and the one on the net position stays insigni�cant.
The key result is thus robust to the exclusion of potentially endogenous loan controls.
A key concern at this stage is that banks also have means for risk management
other than through credit derivatives. Use of these means is conceivably correlated
with credit derivative use. The gross credit derivative position may hence also proxy
for general sophistication in bank risk management. In this case, our estimated e¤ects
cannot (exclusively) be attributed to credit derivatives. To address this issue, regression
5 controls for the stock of other derivatives used for hedging (these derivatives include
interest rate, foreign exchange, equity, and commodity derivatives). The coe¢ cient
on the gross position is essentially unchanged and the other derivatives are turn out
insigni�cant. We have also estimated a version of regression 5 where instead of including
the sum of all other derivatives we include each derivative separately. The results for
our variables of interest are essentially unchanged (not reported here). This suggests
that the risk management bene�ts indeed come through credit derivatives. Among
the other derivatives all are insigni�cant except the commodity derivatives (which are
signi�cant at the 10% level).
Another important issue is the potential endogeneity of the gross credit derivative
position. A bank that pursues a risky strategy may simultaneously underprice in the
11
syndicated lending market and write protection in the CDS market. Alternatively, a
bank that faces good lending opportunities may have low lending rates and hedge the
additional amount of loans using credit derivatives. However, this endogeneity a¤ects
the net position of credit derivatives. It is more di¢ cult to conceive how endogeneity
may a¤ect gross positions. Endogeneity problems are also limited in our setting since
we control for bank �xed e¤ects and time e¤ect. Nonetheless, we also employ an
IV-estimation to account for remaining endogeneity. Our instruments for the gross
position are other derivatives held for trading purposes. Banks typically start hedging
activities in derivatives following trading in derivatives. We thus expect derivatives
for trading to be a good explanatory variable for credit derivatives (Minton, Stulz and
Williamson, 2009, �nd that use of credit derivatives is highly correlated with the trade
of other derivatives). At the same time, we do not expect trading of derivatives to
have a direct independent e¤ect on the lending business of banks: trading is typically
done in response to short-term pro�t opportunities and it di¢ cult to conceive how this
should a¤ect a bank�s lending strategy (in addition in most banks trading activities
and lending activities are carried out in separate organizational entities that do not
communicate). Regression 6 reports results from an IV-regression where the gross credit
derivative position is instrumented with various derivatives held for trading (interest
rate, foreign exchange, equity and commodity derivatives). The F-test of 613.08 in the
�rst stage of the IV regression indicates that trading derivatives are good instruments
as they are highly related with credit derivatives. The J-test has a p-value of 0.35.
This indicates absence of endogeneity for the instruments, con�rming that non-credit
derivatives trading activities are not related to loan pricing. The coe¢ cient of the gross
position is still signi�cant (now only at the 5% level). The size of the coe¢ cient �as
expected �decreases in absolute size, but only slightly (to -9.817).
Another dimension of endogeneity may arise from a contemporaneous dependence
of gross positions on demand or supply side considerations. In regression 7 we thus
include the one-year lagged gross position �instead of the contemporaneous one. The
coe¢ cient now increases in absolute size (to -12.17) and is signi�cant at the 1% level.
We have also carried out various other robustness checks (not reported here), such as
clustering at the �rm level and scaling variables by loans instead of assets, without any
noteworthy change in our variables of interest. We conclude that our results do not
seem to be driven by endogeneity of credit derivative gross positions.
In sum, the evidence in this section suggests a stable negative association between
banks�gross credit derivative positions and loan spreads. The e¤ect is robust to control-
ling for various forms of biases that may arise in this context. No association between
12
net positions and loan spreads can be found. The results thus lend support for the
hypothesis that banks use credit derivatives to manage risks more e¤ectively and pass
on gains to borrowers. By contrast, there is no support for other channels through
which credit derivative may a¤ect loan spreads.
4.3 Loan spreads by borrower type
The baseline analysis shows that borrowers at banks active in credit derivatives bene�t
from lower loan spreads. In this section we analyze whether this e¤ect is uniform across
borrowers, or whether speci�c types of borrowers bene�t more. Since the universe of
liquid credit derivatives mainly consists of large, investment grade corporate borrowers,
we would expect risk management gains to be the largest for these �rms.
For this we add interaction terms between gross positions and borrower types to the
baseline model. Table 4 reports the results. Regression 1 shows the results of a speci�-
cation that looks at whether the credit derivative e¤ect is di¤erent for large �rms. The
dummy variable Large indicates whether a �rm belongs to the 25% largest percentile
of our sample in terms of assets. The interaction term of this variable with the gross
amount in credit derivatives captures the di¤erence in the e¤ect of risk management for
these �rms. The coe¢ cient of the interaction term is negative and signi�cant, indicating
that the largest �rms bene�t more from risk management at banks.
Next, we analyze whether the e¤ect di¤ers between rated and unrated �rms. Rated
�rms are more likely to have liquid credit derivatives given the greater availability of
credit risk information for these �rms. Regression 2 includes an interaction term of the
rating variable with the gross position. As expected �the interaction term is negative
and signi�cant. The risk management bene�t is thus larger for rated �rms. We also
note that the size of the coe¢ cient is large in absolute terms (-19.73). Hence, rated
�rms seem to be a main bene�ciary from bank credit derivative use.
Regression 3 studies whether investment grade �rms experience a di¤erent loan
spread e¤ect. We include interaction terms with dummies indicating whether the �rm
is a low risk entity (i.e., the S&P rating of its senior debt is A or better) or a high risk
entity (i.e., the S&P rating is BBB or worse). The omitted category are unrated �rms.
The low risk interaction term obtains a very high coe¢ cient in absolute values (-42.51)
but is only weakly signi�cant. The low signi�cance most likely re�ects limited rating
coverage in our sample (low risk �rms represent only a fraction of 0.7% in the sample
while high-risk �rms are 16%; the remaining 83.3% are unrated �rms). The combined
coe¢ cient from the interaction term and the non-interacted gross position is -52.87.
13
Thus, a one-standard deviation increase in gross positions at banks results in a loan
spread for �rms rated low-risk that is 44 bps lower (equivalent to a spread reduction of
18%).
We also study whether �rms listed at the stock market bene�t more from banks�use
of credit derivatives. Stock market listing �after controlling for the presence of a rating
� is likely to be unrelated to a �rm�s presence and liquidity in the credit derivative
market. Consistent with this we �nd that the interaction term of stock market listing
and the gross credit derivative position is insigni�cant (see regression 5)
Regressions 1-4 have considered whether �rms more likely to be actively traded
experience di¤erent credit derivative e¤ects. In the respective regressions, the non-
interacted gross-position coe¢ cient stayed signi�cant. This suggests that also �rms less
likely to be actively traded bene�t from enhanced risk management. In regression 6
we address this question directly. We constrain our sample to the set of �rms that
are unrated (and hence are very unlikely to have active credit derivatives trading).
The e¤ect on the gross position is signi�cant and the coe¢ cient (-10.42) is of similar
magnitude as the one in the baseline model. Thus, risk management bene�ts also extend
to the �rms for which the bank cannot directly manage risks using credit derivatives.
This is consistent with risk management (balancing risks within the portfolio, keeping
total risks close to the desired levels and improved measurement of risks) reducing the
banks�overall (marginal) cost of taking on risk. It may also partially re�ect pseudo-
hedging �the practice of banks to hedge untraded exposures using correlated traded
exposures �which allows banks to reduce risks on exposures for which credit derivatives
do not exist.
In sum, the evidence in this section suggests that the �rms that are more likely to be
actively traded in the credit derivative market are the largest bene�ciaries from credit
derivative use at banks.
4.4 Loan spreads during the crisis of 2007-2009
It has been argued that �nancial innovations, while bene�cial in normal times, may
amplify the e¤ects of crises. While this is likely to be the case under (for example) the
incentive channel, the presence of a risk management channel suggests that bene�ts
continue to be present under adverse circumstances. In this section we investigate
whether the di¤erence in loan pricing between active and passive banks persists during
the crisis of 2007-2009. For this purpose, we re-estimate the baseline model allowing
the coe¢ cient of interest and the intercept to di¤er after the onset of the �nancial crisis.
14
Table 5 presents the results. Regression 1 includes a dummy indicating the crisis
period (which we take to start in the last quarter of 2007). This dummy is signi�cant
and its coe¢ cient indicates that loans spreads increase during the crisis period by
42.28 bps. Regression 2 includes the interaction term between the gross position of
credit derivatives and the crisis dummy. The non-interacted gross position term stays
signi�cant and obtains a coe¢ cient of -12.29. The interacted gross position term is
insigni�cant. This suggests that the bene�ts of credit derivative use remain unchanged
after the onset of the �nancial crisis.
A concern with regression 2 is that banks may have changed their credit derivative
activities in response to the crisis. The crisis interaction term in regression 2 relates to
the contemporaneous gross position. It thus does not directly measure bene�ts from
risk management prior to the crisis. In regression 3 we look at how loan spreads change
for banks depending on their credit derivative active prior to the crisis. We thus include
an interaction term of the crisis dummy with banks�gross position in the third quarter
of 2007. We �nd that the interaction term remains negative and insigni�cant. The
persistence of the loan spread bene�t is thus not driven by banks� responses to the
crisis but by prior engagement in credit derivative markets.
We �nally consider whether net positions in credit derivative markets lead to dif-
ferent loan spreads in the crisis. We thus include the net position and the net position
interacted with the crisis dummy. The interaction term is insigni�cant. We also note
that our prior results are unchanged as the non-interacted net term remains insigni�cant
as well.
In conclusion, the evidence suggests that even though loan spreads generally in-
creased after the onset of the �nancial crisis, the bene�ts of borrowing from banks�
engaging in risk management via credit derivatives persist during the crisis.
4.5 Credit derivative use and bank lending
The evidence from the loan-level regressions supports the hypothesis that banks use
credit derivatives for risk management purposes. In this section we look at banks�
lending characteristics in general. If banks successfully manage their risks, we would
expect banks active in credit derivative markets to experience lower losses on loans. In
addition, we would expect these banks to be less likely to be constrained when credit
risks in the economy worsen and exhibit a more stable lending behavior.# 8
8Figure 2 already suggested that the loan pricing behavior of active banks is more stable than the
one of passive banks (the standard deviation of quarterly spreads of the CDS banks are #, compared
15
Speci�cally, we relate in this section lending characteristics at the bank level to
their use of credit derivatives. First, we study whether charge-o¤s of commercial and
industrial loans are related to credit derivative use and whether this e¤ect changes
during the crisis. Second, we compare the lending volume of active and passive banks
before and during the crisis. For this analysis we use bank level data from the Call
Reports. We include in our sample observations for the years 2006 to 2010. We estimate
two models:
Netchargeoffs=TAb;t = �+ �1Crisist + �2GrossCDb;t + �3Crisist �GrossCDb;t
+
KXi=1
�iBi;b;t + �b;t (2)
CommercialLoans=TAb;t = �+ �1Crisist + �2GrossCDb;t + �3Crisist �GrossCDb;t
+KXi=1
�iBi;b;t + �b;t (3)
In the �rst model, the dependent variable is the sum of net charge-o¤s (charge-o¤s minus
recoveries) of commercial and industrial loans minus the net gains of credit derivatives
scaled by assets. We include the gains on credit derivatives in order to capture potential
risk management bene�ts: if a bank e¤ectively manages its risk, charge-o¤s (recoveries)
of loans should be o¤-set by gains (losses) in credit derivatives holdings. The term
B stands for other bank characteristics. These include: subordinated debt, equity,
liquid assets, total loans and commercial loans (scaled by assets). We also include the
logarithm of assets and the ROA.
If credit derivative use extends risk management bene�ts, we should see that banks
with larger gross amounts of credit derivatives face a lower level of net charge-o¤s in a
given period. We hence expect the coe¢ cient on the gross amount of credit derivatives
to be negative in the �rst model. The crisis regressions have shown that (although
spreads increased across the board) the loan spread di¤erential between active and
passive banks persisted during the crisis. This suggests that banks with active risk
management were not more constrained by loan losses than other banks. Accordingly,
we expect the interaction term of the gross position and the crisis dummy in the model
to be insigni�cant.
The dependent variable in the second model are commercial loans scaled by assets.
We include the same set of bank controls but exclude the dependent variable. Banks
that successfully manage their risk should be less constrained under adverse conditions.
to # for the non-CDS banks.
16
They should have more stable lending and possibly be able to expand lending activities
(relative to passive banks) in crises times. We thus expect the interaction term of the
gross derivative position with commercial lending to be non-negative or even positive.
Table 6 displays the results of both models. In both regressions standard errors
are clustered at bank level. Regression 1 displays the results for the net charge-o¤
regression. We see that active banks have signi�cantly lower charge-o¤s as indicated
by the coe¢ cient of the gross positions. The coe¢ cient on the crisis dummy is positive
and signi�cant �indicating that charge-o¤s increased during the crisis. The interaction
term of the crisis dummy with the gross position is insigni�cant. Thus, the advantage
of active banks (in terms of lower charge-o¤s) persists during the crisis.
Regression 2 estimates the lending volume model. We �nd that the coe¢ cient for
the gross position in credit derivatives is not signi�cant in this regression, indicating
that active users of credit derivatives do not extend more commercial and industrial
loans than other banks. The negative sign on the crisis dummy shows that the volume
of commercial and industrial loans extended by banks overall decreases during the
crisis. The interaction terms of the crisis dummy and the gross position is positive and
signi�cant. Thus, banks active in credit derivatives markets increased their lending
volume relative to passive banks. This is consistent with risk management stabilizing
the lending activities of these banks.
Summarizing, the bank-level regressions suggest that banks active in credit deriva-
tive markets face lower charge-o¤s in both normal and crises times. In addition, they
are able to expand their lending relative to passive banks in crisis times. These �ndings
are consistent with risk management bene�ts from credit derivative use.
5 Conclusions
The debate on the role of �nancial innovations is still ongoing. There is no consensus
about whether their impact on the �nancial system is broadly a positive one or not �
and what the sources of their e¤ects are. In this paper we try to learn about �nancial
innovations and their role for the economy by studying their impact on loan pricing. We
do this by focusing on credit derivatives �probably the most signi�cant �nancial inno-
vation of the recent decade. There are several channels through which credit derivatives
can impact lending behavior (and thus a¤ect economy activity). We derive hypotheses
that relate di¤erent uses of credit derivatives to loan spreads and derive a new empirical
strategy that allows us to identify the channel through which the e¤ect occurs.
Based on matched data from the LPC Dealscan database and Call Reports, we
17
estimate a loan pricing model that controls for loan, borrower and bank characteristics.
Our key result is that banks� gross position in credit derivatives has a signi�cantly
negative and robust e¤ect on corporate loan spreads. We argue that this indicates that
banks use credit derivatives for risk management and pass the arising bene�ts (at least
partly) on to borrowers. Such bene�ts include a better risk-balance in the loan portfolio,
an improved ability to keep risk-levels at target ratios but also banks generally getting
more sophisticated about the measurement and control of credit risks. We also �nd that
the bene�ts from risk management persist after the onset of the �nancial crisis. Banks
that actively manage their risks with credit derivatives exhibit also lower losses and
an more stable supply of loans during the �nancial crisis. Taken together, our paper
provides consistent evidence on bene�cial real e¤ects of �nancial innovations that are
present independently of economic conditions.
18
References
[1] Adam, T., Guettler, A. (2010) The Use of Credit Default Swaps by U.S. Fixed-
Income Mutual Funds. Working Paper, July 2010.
[2] Ashcraft, A. and Santos, J. (2009) Has the CDS Market Lowered the Cost of
Corporate Debt, Journal of Monetary Economics 56, 514-523
[3] BIS (2005) Credit Risk Transfer, The Joint Forum Bank for International Settle-
ments
[4] Berger, A., Frame, S., and Ioannidou, V. (2011) Tests of Ex Ante Versus Ex Post
Theories of Collateral Using Private and Public Information, Journal of Financial
Economics 100, 85-97
[5] Berger, A. and Udell, G.(1990) Collateral, loan quality, and bank risk, Journal of
Monetary Economics 25, 21-42
[6] Besanko, D. and Thakor, A. (1987) Collateral and Rationing: Sorting equilibria in
monopolistic and competitive credit markets, International Economic Review 28,
671-689
[7] Blanco, R., Brennan, S., and Marsh, I. (2005) An empirical analysis of the dy-
namic relation between investment-grade bonds and credit default swaps,Journal
of Finance 60, 2255-2281
[8] British Banker�s Association (2006) Credit Derivatives Report
[9] Cebenoyan, A. and Strahan, P. (2004) Risk Management, Capital Structure and
Lending at Banks, Journal of Banking and Finance 28, 19-43
[10] Chan, Y. and Kanatas, G., (1985) Asymmetric Valuations and the Role of Collat-
eral in Loan Agreements,Journal of Money, Credit and Banking 17, 84-95
[11] Dennis, S. A. Nandy, D. and Sharpe, I. (2000), The determinants of contract
terms in bank revolving credit agreements, Journal of Financial and Quantitative
Analysis 35, 87-110
[12] Du¢ e, D. (2008) Innovations in credit risk transfer: Implications for �nancial
stability, BIS Working paper No. 255
19
[13] Ellul, A., Yerramilli, V. (2010) Stronger Risk Controls, Lower Risk: Evidence from
U.S. Bank Holding Companies, 2011 AFA Denver Meetings Paper
[14] Franke, G. and Krahnen, J. P. (2005) Default Risk Sharing between Banks and
Markets: The Constribution of Collateralized Debt Obligations, CFS Working
Paper Series No. 2005/06
[15] Froot, K., Scharfstein, D. and Stein, J. (1993) Risk Management: Coordinating
Corporate Investment and Financing Policies, Journal of Finance 48, 1629-1658
[16] Goderis, B., Marsh, I., Vall Castello, J. and Wagner, W.(2007) Bank behaviour
with access to credit risk transfer markets, Research Discussion Papers 4/2007,
Bank of Finland
[17] Guener, B., (2006) Loan Sales and the Cost of Corporate Borrowing, The Review
of Financial Studies 19, No. 2
[18] Harjoto, M., Mullineax, D.J., and Ha-Chin Yi (2006) A Comparision of Syn-
dicated Loan Pricing at Investment and Commercial Banks, Financial Manage-
ment;35,4;ABI/INFORM Global pg.49
[19] Hirtle, B., (2009) Credit Derivatives and Bank Credit Supply, Journal of Financial
Intermediation 18(2), 125-150
[20] Loutskina, E. and Strahan, P., (2006) Securitization and the declining impact of
bank �nance on loan supply: Evidence from mortgage acceptance rates, Journal
of Finance 64, 861-889
[21] Marsh, I. (2006) The E¤ect of Lenders�Credit Risk Transfer Activities on Borrow-
ing Firms�Equity Returns, Working paper, Cass Business School.
[22] Minton, B., Stulz, R. and Williamson, R. (2009) How Much Do Bank Use Credit
Derivatives to Hedge Loans?, Journal of Financial Services Research 35, 1-31
[23] Morrison, A. (2005) Credit Derivatives, Disintermediation and Investment Deci-
sions, Journal of Business 78(2), 621-647
[24] Nadauld, T. and Weisbach, M. (2011) Did Securitization a¤ect the Cost of corpo-
rate Debt?, NBER Working paper No. 16849
[25] Norden, L. and Wagner, W. (2008) Credit derivatives and loan pricing, Journal of
Banking and Finance 32(12), 2560-2569
20
[26] Norden, L. and Weber, M. (2004) Informational e¢ ciency of credit default swap
and stock markets: The impact of credit rating announcements, Journal of Banking
and Finance 28(11), 2813-2843
[27] Strahan, P. (1999) Borrower risk and the price and nonprice terms of bank loans,
Working paper, Banking Studies Function, Federal Reserve Bank of New York
[28] Stulz, R. (2010) Credit Default Swaps and the Credit Crisis,Journal of Economic
Perspectives, American Economic Association 24(1), 73-92
21
6 Appendix
22
Table 2: Descriptive Statistics
Variables Mean Standard Deviation Minimum Maximum
Loan Characteristics
Spread 238.179 105.646 30 455
Log(amount) 17.828 1.318 10.665 22.86
Secured 0.338 0.473 0 1
Unsecured 0.038 0.192 0 1
Short Maturity 0.174 0.379 0 1
Intermediate Maturity 0.490 0.499 0 1
Long Maturity 0.334 0.471 0 1
TERM 0.736 0.440 0 1
TERM A 0.090 0.287 0 1
TERM B 0.160 0.366 0 1
TERM C 0.013 0.114 0 1
Borrower Characteristics
Log(sales) 19.006 1.950 9.452 21.581
Ticker 0.265 0.441 0 1
Rating 0.016 0.126 0 1
AAA 0.0001 0.011 0 1
AA 0.001 0.040 0 1
A 0.006 0.078 0 1
BBB 0.024 0.154 0 1
BB 0.052 0.222 0 1
B 0.073 0.261 0 1
CCC 0.011 0.106 0 1
CC 0.0007 0.027 0 1
C 0 0 0 0
Bank Characteristics
Gross CD/TA 0.453 0.823 0 3.988
Net CD/TA 0.023 0.055 -0.039 0.225
Derivatives not for trade/TA 0.273 0.310 0 1.263
Log(assets) 19.006 1.950 9.452 21.581
ROA 0.005 0.006 -0.073 0.068
Sub Debt/TA 0.333 0.146 0.0006 0.848
Liquid Assets/TA 0.177 0.096 0 0.991
Charge-o¤s/TA 0.003 0.004 0 0.072
Equity/TA 0.085 0.072 0.010 0.995
23
Table 3: Credit derivative use and loan spreads
(1) (2) (3) (4) (5) (6) (7)
Variable Spread Spread Spread Spread Spread Spread Spread
Gross CD/TA -9.362*** -10.93*** -14.58*** -10.75*** -9.817**
(2.523) (2.146) (2.032) (2.120) (4.576)
Net CD/TA 38.12 13.12 -1.285 17.34 12.39 11.66
(44.92) (29.19) (28.27) (27.65) (29.16) (45.09)
Derivatives 2.602
not for trade/TA (9.008)
Gross CD/TA lag -12.17***
(3.735)
Net CD/TA lag -9.598
(22.22)
Log(sales) -4.483*** -4.356*** -4.292*** -9.974*** -4.358*** -4.350*** -3.804**
(1.483) (1.441) (1.446) (2.184) (1.440) (1.554) (1.566)
AAA -9.466 -14.78 -21.85* -11.91 -15.37 -15.50
(10.32) (11.94) (12.03) (10.62) (11.61) (16.86)
AA 27.40 15.74 13.89 44.93 15.32 15.55 26.64
(83.12) (86.90) (86.13) (90.79) (86.87) (83.01) (90.80)
A -11.77 -9.988 -10.71 -4.787 -10.15 -10.06 -0.835
(35.81) (36.16) (37.33) (35.80) (35.99) (27.04) (39.23)
BB 38.38*** 39.32*** 38.85*** 72.41*** 39.36*** 39.27*** 43.07***
(7.056) (7.133) (7.244) (6.275) (7.105) (9.635) (9.965)
B 78.43*** 79.46*** 79.07*** 119.1*** 79.48*** 79.42*** 83.29***
(8.890) (9.061) (8.906) (9.073) (9.048) (9.646) (11.77)
CCC 129.1*** 129.4*** 128.8*** 170.9*** 129.5*** 129.4*** 143.0***
(10.48) (10.79) (10.92) (9.897) (10.69) (14.24) (14.34)
CC 223.3*** 224.7*** 226.1*** 252.6*** 224.0*** 224.8*** 226.5***
(47.49) (46.03) (45.10) (40.60) (46.36) (41.19) (46.89)
NR 55.74*** 56.20*** 55.32*** 79.96*** 56.25*** 56.11*** 59.21***
(6.792) (6.874) (6.865) (6.410) (6.848) (8.957) (10.40)
Rating -28.83* -28.48* -28.94* -29.92*** -28.37* -28.53* -33.64*
(14.70) (15.09) (15.52) (10.96) (14.94) (15.55) (17.47)
Ticker -9.873 -6.318 -6.814 -6.993 -6.325 -6.369 -8.695
(7.271) (6.919) (6.952) (7.435) (6.920) (4.210) (6.980)
Log(amount) -13.28*** -14.58*** -14.51*** -14.58*** -14.57*** -13.64***
(2.406) (2.483) (2.467) (2.488) (2.188) (3.045)
24
Table 3: Credit derivative use and loan spreads (cont.)
(1) (2) (3) (4) (5) (6) (7)
Variables Spread Spread Spread Spread Spread Spread Spread
Secured 15.28*** 14.70*** 14.43*** 14.73*** 14.67*** 10.57*
(5.267) (5.063) (5.111) (5.041) (4.491) (5.389)
Unsecured -56.08*** -55.30*** -56.24*** -55.27*** -55.40*** -54.35***
(7.696) (8.279) (8.385) (8.304) (8.135) (9.450)
Interm. maturity -11.61 -13.19 -13.70 -13.18 -13.24** -12.95*
(9.062) (9.001) (9.076) (8.982) (5.826) (7.761)
Long maturity -14.20* -11.59 -11.87 -11.57 -11.62* -11.85
(7.968) (7.559) (7.608) (7.553) (6.453) (9.096)
TERM A 26.58*** 23.81*** 24.40*** 23.78*** 23.87*** 23.36***
(5.657) (4.951) (5.092) (4.943) (5.710) (5.603)
TERM B 58.38*** 53.90*** 54.65*** 53.89*** 53.98*** 51.81***
(6.144) (6.611) (6.602) (6.604) (5.331) (7.108)
TERM C 45.98*** 39.61*** 41.46*** 39.65*** 39.80*** 31.78***
(10.10) (9.648) (9.452) (9.650) (10.15) (10.37)
ROA -351.2
(309.0)
Subdebt/TA 1.639
(30.43)
Liquid Assets/TA -34.77
(24.92)
Chargeo¤/TA 1,648***
(464.5)
Log(assets) -4.135
(2.494)
Equity/TA -9.369
(26.14)
F-stat IV 613.08
J-test p-value 0.356
Industry Dummies Yes Yes Yes Yes Yes Yes Yes
Purpose Dummies Yes Yes Yes No Yes Yes Yes
Year Dummies Yes Yes Yes Yes Yes Yes Yes
Bank Fixed E¤ects No Yes Yes Yes Yes Yes Yes
Observations 2,559 2,638 2,638 2,638 2,638 2,638 2,322
R-squared 0.362 0.398 0.396 0.330 0.398 0.398 0.385
The dependent variable is the loan spread spread over LIBOR (basis points). All models are estimated using OLS with
clustered robust standard errors at the bank level (in parentheses). ***, ** and * denote signi�cance at the 1%, 5% and
10% level respectively. Equation in column 5 uses IV estimation.
25
Table 4: Loan spreads by borrower type
(1) (2) (3) (4) (5)
Variables Spread Spread Spread Spread Spread
Gross CD/TA -6.890*** -10.05*** -10.36*** -10.62*** -10.43***
(1.954) (1.902) (2.385) (3.689) (1.954)
Large -10.55
(8.848)
Gross CD/TA*large -7.397***
(2.343)
Rating -27.08* -18.55 -28.48*
(14.42) (14.37) (15.18)
Gross CD/TA*rating -19.73***
(5.991)
Low_risk -46.31
(43.95)
High_risk 2.174
(3.689)
Gross CD/TA*high_risk -0.117
(2.013)
Gross CD/TA*low_risk -42.51*
(23.11)
Ticker -5.586 -6.419 -10.34 -6.143 -8.532
(7.142) (6.959) (6.345) (7.355) (6.852)
Gross CD/TA*ticker -0.309
(4.100)
Borrower Controls Yes Yes Yes Yes Yes
Loan Controls Yes Yes Yes Yes Yes
Year Dummies Yes Yes Yes Yes Yes
Bank Fixed E¤ects Yes Yes Yes Yes Yes
Observations 2,638 2,638 2,638 2,638 2581
R-squared 0.400 0.399 0.363 0.398 0.353
The dependent variable is the loan spread spread over LIBOR (basis points). All
models are estimated using OLS with clustered robust standard errors at the bank
level (in parentheses). ***, ** and * denote signi�cance at the 1%, 5% and 10% level
respectively. Model (5) only includes non rated �rms
26
Table 5: Loan spreads during the crisis of 2007-2009
(1) (2) (3) (4)
Variables Spread Spread Spread Spread
Crisis 42.28*** 44.21*** 44.45*** 44.49***
(13.70) (14.28) (13.52) (14.31)
Gross CD/TA -12.29*** -12.23*** -12.47***
(2.007) (1.967) (2.185)
Gross CD/TA*crisis 0.325 -2.544
(3.307) (4.824)
Net CD/TA 20.83
(26.57)
Net CD/TA*crisis 136.2
(164.7)
Gross CD 07/TA*crisis 0.115
(2.530)
Borrower Controls Yes Yes Yes Yes
Loan Controls Yes Yes Yes Yes
Year Dummies Yes Yes Yes Yes
Bank Fixed E¤ects Yes Yes Yes Yes
Observations 4,022 2,596 2,596 2,596
R-squared 0.417 0.389 0.389 0.389
The dependent variable is the loan spread spread over LIBOR (basis points). All
models are estimated using OLS with clustered robust standard errors at the bank
level (in parentheses). ***, ** and * denote signi�cance at the 1%, 5% and 10% level
respectively.
27
Table 6: CDS use and Bank Lending
(1) (2)
Variables Charge-o¤s commercial/TA Commercial loans/TA
Crisis 0.0003*** -0.032**
(4.95e-05) (0.012)
Gross CD/TA -0.259** -7.783
(0.103) (19.62)
Gross CD/TA*crisis 0.115 52.44***
(0.115) (20.29)
Sub debt/TA 4.23e-05 -0.057**
(0.0001) (0.026)
Liquid assets/TA 0.0004*** 0.094***
(0.0001) (0.028)
Equity/TA 0.0006*** 0.032
(0.0002) (0.036)
Log(assets) 8.88e-05*** 0.008***
(1.36e-05) (0.001)
Total loan/TA 0.0008*** 0.220***
(0.0001) (0.028)
Commercial loans/TA 0.003***
(0.0002)
ROA -0.034*** 0.208
(0.001) (0.203)
Constant -0.002*** -0.154***
(0.0002) (0.042)
Observations 2,243 2,243
R-squared 0.355 0.138
The dependent variable in these models are: In model (1) Net charge-o¤s minus CDS
gains scaled by total assets. In model (2) total volume of commercial loan extended
scaled by total assets. All models are estimated using OLS with clustered robust
standard errors at the bank level (in parentheses). ***, ** and * denote signi�cance
at the 1%, 5% and 10% level respectively.
28
Figure 1: Evolution Spreads CRT vs non CRT Banks
Figure 2: Evolution Gross Credit derivatives position
Figure 3: Evolution Net Credit derivatives position
29